what makes the PL cursor life-cycle must be in the same transaction?

begin;

declare cur cursor for select * from t;

insert into t2 values(...);

fetch next cur;

commit;

// after this,  I can't fetch cur any more. 

My question are:

1.  Is this must in principle?  or it is easy to implement as this in PG?

2.  Any bad thing would happen if I keep the named portal (for the cursor) available even the transaction is commit, so that I can fetch the cursor after the transaction is committed?  

Thanks